From Blanket to Precision: A Data-Driven Decision Support Tool for Site-Specific Fertilizer Recommendations in Ethiopia

Ethiopia’s smallholder farmers continue to face large yield gaps despite decades of fertiliser use and agricultural research investment. A key constraint has been blanket fertiliser recommendations: uniform rates applied nationwide without accounting for local soils, agro‑ecological conditions, or crop diversity. This one‑size‑fits‑all approach leads to systematic misapplication, with over‑use in some areas and under‑use in others. As a result, Ethiopian wheat farmers achieve only 35-45% of potential yields, while more than 60 years of agronomic research data remained fragmented. The Supporting Soil Health Interventions in Ethiopia (SSHI) project, co‑funded by BMZ and the Gates Foundation and implemented by GIZ with the Alliance of Bioversity International and CIAT, EIAR, and the Ministry of Agriculture, addressed this gap by developing Ethiopia’s first Site‑Specific Fertiliser Decision Support Tool. Using machine learning, the tool delivers localised fertiliser recommendations, validated across hundreds of farms.